EEmc Gammas via conversion method, retool

A few notes on previous studies.

1.0 Definition of D=log(Epost/Esmd) leads to "underflow" events when Epost=0.

Hal pointed out that the EEmc Gammas via conversion method, systematics II using postshower information has a flaw.  It drops the class of events where there was no energy deposited w/in the postshower detector.  i.e. it loses the ~2k events where the hadronic veto cut would have been satisfied.  See EEmc Gammas via conversion method on the conversion method.  So it will be necessary to divide the events into two classes.

Class I: Events with energy deposited in the postshower layer.  Use the EEmc Gammas via conversion method, systematics II and correct for the hadronic background.

Class II: Events with NO energy deposited in the postshower layer.  Previously we have seen discrepances EEmc Gammas via conversion method which are suggestive of hadronic contamination.  My guess is that an event sample with Epost == 0, as opposed to the original Epost<0.2 MeV, will be pretty much clean.  But this should be cross checked somehow. 

2.0 Fill 7863 appears to be missing in some or all of the studies.

3.0 Luminosity was likely calculated incorrectly in the pythia studies.  The manual states that "nTried" is only present for a cross-check of the internals of pythia.  The nAccepted or selected values should be compared w/ the cross section.  This leads to a ~x3 overestimate of luminosities for a given event sample.

4.0 Trigger efficiencies near threshold have not been taken into account previously.  This leads to an underestimate of the cross section / total number of gammas for a given luminosity.

In fact, there were multiple problems with this MC sample.  pT was thrown starting at 5 GeV.  Trigger turns on at 5.2 GeV.  No events were available to
migrate to higher pT... which leads to underestimate of the efficiency.  Vertex was too narrow.  Also leads to an underestimate of the efficiency.